摘要 |
<p>PROBLEM TO BE SOLVED: To decide a suitable operational action and manipulated variable for a short time in order to shift a process state into a target state and maintain without depending on the experience of an operator. SOLUTION: Based on the latest actual values of the operational action and the manipulated variable to a blast furnace 15 and the detecting end measurement information of the blast furnace 15, a corrective action and manipulated corrective variable for correcting the difference between a controlling amount predicted by a state predicting part 3 and a target value, are inferred by a correlating rule in an inference processing part 9 and a correcting amount making part 11. Then, a convergent judgement of the manipulated corrective variable is performed in a convergent condition judging part 13. According to the judged result, the manipulated corrective variable is added to the above manipulated variable so as to supply the value into the blast furnace 15 from an operational control part 14 as the next operational action and manipulated variable. Then, after performing this operation to the blast furnace 15, the suitable operational action and the manipulated variable can be inferred and decided without waiting for long time until a response is received.</p> |